VOLUME 26 NUMBER 5
M A Y 1993 Registered in U.S. Patent and Trademark Office; Copyright 1993 by the American Chemical Society
Solution Structures of Lithium Dialkylamides and Related N-Lithiated Species: Results from 6Li-15NDouble Labeling Experiments DAVIDB. COLLUM Department of Chemistry, Baker Laboratory, Cornell University, Zthaca, New York 14853-1301 Received August 18, 1992 (Resubmitted Manuscript Received J a n u a r y 12, 1993)
The chemistry of lithium is pervasive within organic chemistry.’ It would appear, for example, that well over 95% of natural products syntheses rely upon lithium-based reagenh in one form or another. At the same time, however, the coordination chemistry of organolithium reagents and the mechanisms of organolithium-based reactions are extraordinarily complex and poorly u n d e r s t o ~ d .This ~ ~ ~complexity stems from a high propensity of organolithium compounds to selfassociate into higher aggregates, a phenomenon that is markedly dependent on the choice of solvent and precise conditions. The problem is intensified by the rapid subunit and solvent exchanges within the various aggregates as well as the extreme oxygen and moisture sensitivity of most organolithium derivatives. As a result, a vast number of organolithium reagents are routinely generated and used without direct evidence of their solution structures, dynamic behavior, or even existence. Lithium dialkylamidesand related N-lithiated species constitute a very important class of organolithium reagents. They are the preferred bases for formation of ketone enolates114and other stabilized carbanions as well as for generating low steady-state concentrations of some relatively unstable carbani0ns.l~~ The lithioimines and related lithiated Schiff s bases have played prominently as ketone enolate equivalents in cases where the enolates have proven less than adequate.6 In DavM Collum received a bachelor’s degree in biology from the Cornell University College of Agriculture and Life Sciences in 1977. After receiving a Ph.D in 1980 from Columbia University worklng with Clark Still, he returned to the Department of Chemistry at Cornell, where he is now a Professor of Chemistry. His work at Cornellhasaddressedtopics In naturalproductssynthesis, organotransitlonmetal Chemistry,and organolithiumstructure and mechanism.
fact, the progress in synthetic organic chemistry over the last three decades would certainly have been less impressive without the contributions from lithium amide chemistry to the refinement of carbon-carbon bond forming technology. This Account describes the solution structures of lithium dialkylamides and closely related N-lithiated species that have been determined by 6Liand 15NNMR spectroscopic analyses of isotopically enriched substrates. Strategies are described for extracting structural details from both inordinately simple and inordinately complex NMR spectra. Since a logical, stepwise presentation and a chronologicalnarrative are mutually exclusive,the former will be attempted under pretense of the latter. (1)(a) Seebach, D.Angew. Chem., Int. Ed. Engl. 1988,27,1624. (b) Stowell, J. C. Carbanions in Organic Synthesis; Wiley: New York, 1979. (c) Asymmetric Synthesis; Morrison, J. D., Ed.; Academic Press: New York, 1983;Vols. 2 and 3. (2)Ions and Ion Pairs in Organic Reactions; Szwarc, M., Ed.; Wiley: New York, 1972;Vols. 1 and 2. (3)(a) Klumpp, G.W. Recl. Trau. Chim. Pays-Bas 1986,105,1. (b) Setzer, W.N.; Schleyer, P. v. R. Adu. Organomet. Chem. 1985,24,354. (c) Fraenkel, G.;Hsu, H.; Su, B. P. In Lithium: Current Applications in Science, Medicine, and Technology; Bach, R. O., Ed.; Wiley: New York, 1985;Chapter 19 and references cited therein. (4)dAngelo, J. Tetrahedron 1976, 32, 2979. Heathcock, C. H. In Comprehensiue Carbanion Chemistry; Buncel, E., Durst, T., Eds.; Elsevier: New York, 1980;Vol. B, Chapter 4. (5)Snieckus, V. Chem. Reu. 1990,90,879. (6)Bergbreiter, D. E.; Momongan, M. In Comprehensiue Organic Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon: New York, 1989; Vol. 1,p 503. Martin, S. F. In Comprehensive Organic Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon: New York, 1989; Vol. 1, p 475. Hickmott, P. W. Tetrahedron 1982,38, 1975. Enders, D. In Current Trends in Organic Synthesis; Nozaki, H., Ed.; Pergamon: New York, 1983. Whitesell, J. K.;Whitesell, M. A. Synthesis 1983,517. Fraser, R. R.In Comprehensiue Carbanion Chemistry; Buncel, E., Durst, T., Eds.; Elsevier: New York, 1980.
0001-4842/93/0126-0227$04.00/00 1993 American Chemical Society
228 Acc. Chem. Res., Vol. 26,No. 5, 1993
Collum
The Problem of Symmetry. Several excellent reviews offer details of lithium amide structure studies that will not be repeated here.7 The problem of symmetry and consequent numbing spectral simplicity is well worth emphasizing, however. One typically finds that NMR spectra of lithium amides contain a single ‘jLi or 7Li resonance and single sets of 13C and ‘H resonances consistent with many plausible structures including free ions, solvent-separated ion pairs, monomers, dimers, cyclic oligomers, and polymers. The reliance on colligative measurements for determining aggregation is not without ita problem^;^ structural details are available only through considerable inference, colligative data are vulnerable to unseen impurities, and conclusions are often difficult to verify by independent methods. Nevertheless, substantial progress has been made in a few instances as best exemplified by the seminal studies of lithium hexamethyldisilazide (LiHMDS) reported by Kimura and Brown in 197L8 They found that LiHMDS features solvent- and concentration-dependent spectroscopic and colligative properties consistent with a monomerdimer mixture in ethereal solventa (eq 1)and a dimertetramer mixture in hydrocarbons (eq 2). In many respects, their progress can be attributed to the disruption of the spectral simplicity due to the coexistence of multiple aggregation states. The success of the ‘jLi-15N double labeling methods described in this Account arises from the disruption of spectral simplicity as well.
* (Me,Si),NLiS,
(1)
[(Me,Si),NLil, + [(Me,Si),NLil,
(2)
[ (Me,Si),NLiS,l,
I. Simple Equilibria Lithioimines and Lithium Anilides. Our interest in the structures of N-lithiated species originated from investigations of lithiated hydrazone alkylationsg and intensified during studies of lithioimine 1.l0 The 13C and ‘jLi spectra of 1 in the presence of low THF concentrations revealed two discrete sets of resonances that could be attributed to a pair of isomeric species in a 2:l ratio. Colligative techniques described in the literature” for determining both aggregation and solvation state provided some assistance. Titration of solvent-free lithiated imine with limited quantities of THF revealed average solution molalities consistent with disolvated dimers. At elevated THF concentrations a third species appeared that is demonstrably more solvated and less aggregated than the putative dimers. The simplicity of the NMR spectra of 1 forced us to turn to 6Li-15N double labeling for more definitive structural insight. The low quadrupole moment of ‘jLi (7) Gregory,K.; Schleyer, P. v. R.; Snaith, R. Ado. Inorg. Chem. 1991, 37,47. Mulvey, R. E. Chem. SOC.Reu. 1991,20, 167. (8) Kimura, B. Y.; Brown, T. L. J. Organomet. Chem. 1971, 26, 57. (9) Wanat, R. A.; Collum, D. B.; Van Duyne, G.; Clardy, J.; DePue, R. T. J . Am. Chem. SOC.1986, 108, 3416. (10) Kallman, N.; Collum, D. B. J. Am. Chem. SOC.1987,109, 7466. (11) Eastham, J. F.; Gibson, G. W. J . Am. Chem. SOC.1963,85, 2171. Cheema, Z. K.; Gibson, G. W.; Eastham, J. F. J. Am. Chem. SOC.1963, 85, 3517. Brown, T. L.; Gerteis, R. L.; Bafus, D. A.; Ladd, J. A. J . Am. Chem. SOC.1964,86,2135. Bartlett, P. D.; Goebel, C. V.; Weber, W. P. J . Am. Chem. SOC.1969, 91, 7425. Lewis, H. L.; Brown, T. L. J . Am. Chem. SOC.1970, 92, 4664. Quirk, R. P.; Kester, D. E. J. Organomet. Chem. 1977,127,111. Arnett, E. M.; Fisher, F. J.; Nichols, M. A.;Ribeiro, A. A. J . Am. Chem. SOC.1990, 112,801.
1
5
and the reasonable sensitivity of both 6Li (spin 1) and 15N (spin l/d suggested that ‘jLi-15N scalar coupling observed by ‘jLiand 15NNMR spectroscopycould afford insight into N-Li connectivities (Table I). Although previous NMR spectroscopic studies of 15N-labeled lithium amides did not uncover coupling at ambient temperatures,’, we suspected that coupling would be observable at low temperatures where aggregatesubunit exchanges typically become slow on the NMR time scales. As we initiated our investigations Professor Lloyd Jackman and co-workers at Penn State were already making substantial progress in their studies of [‘jLi,15N]lithiumanilides (vide infra).13-15 ‘jLi and 15N NMR spectra of 16Li,15N]1display resonance multiplicities in full accord with a 2:l mixture of two isomeric dimers at low THF concentrations and a monomer in neat THF. Supplemented with additional information on the quantitative concentration dependencies of the monomer-dimer ratio, we completed the aggregation-state and solvation-state assignments summarized in Scheme I. Unfortunately, the resonances ascribed to the 2:l mixture of dimers 2 and 3 are equally consistent with a single stereoisomeric trimer 5. Three years later single-frequency decouplings provided the explicit resonance correlations necessary to rigorously exclude trimer 5.’6 Singlefrequency decoupling became an important component of our studies by allowing us to systematically unravel the ‘jLi-15N resonance correlations of complex equilibria. Jackman’s early studies of [6Li,15N]lithiumanilides revealed predominantly monomer-dimer mixtures in analogy to lithiated imines.13J4 The monomer-dimer ratios depend on the precise substitution pattern of the anilide and solvent chosen. Generally, increasing the steric demands of the N-alkyl moieties promotes formation of monomer. In several instances they observed resonances consistent with aggregate stereoisomers analogous to 2 and 3. Their method for determining solvation states is especially n0tew0rthy.l~ They found that ‘Li quadrupolar splitting constants (QSCs) determined from ‘Li and 13C spin-lattice relaxation times correlate with the degree of solvation (12) Mansour, T. S.; Wong, T. C.; Kaiser, E. M. J. Chem. SOC., Perkin Trans. 2 1985, 2045. Ide, S.; Iwasawa, K.; Yoshino, A.; Yoshida, T.; Takahashi, K. Magn. Reson. Chem. 1987,25,675. Konishi, K.; Yoshino, A.; Katoh, M.; Mataumoto, K.; Takahashi, K.; Iwamura, H. Chem. Lett. 1982, 169. Barchiesi, E.; Bradamante, S.J. Phys.Org. Chem. 1990, 3, 139. Chivers, T.; McIntyre, D. D.; Schmidt, K. J.; Vogel, H. J. J. Chem. SOC.,Chem. Commun. 1990,1341. Vaes, J.; Chabanel, M.; Martin, M. L. J . Phys. Chem. 1978,82, 2420. (13) Jackman, L. M. Aduances in Carbanion Chemistry. Presented at the 190thNational Meeting of the American ChemicalSociety,Chicago, IL, 1985. (14) Jackman, L. M.; Scarmoutzos, L. M. J . Am. Chem. SOC.1987,109, 5348. (15) Jackman, L. M.; Scarmoutzos, L. M.; Porter, W. J. Am. Chem. SOC.1987,109,6524. Jackman, L. M.; Scarmoutzos, L. M.; Smith, B. D.; 1988, 110,6058. Williard, P. G. J . Am. Chem. SOC. (16) Gilchrist, J. H.; Harrison,A. T.;Fuller, D. J.; Collum, D. B. J. Am. Chem. SOC.1990, 112, 4069. (17) Jackman, L. M.; Scarmoutzos, L. M.; DeBrosse, C. W. J . Am. Chem. SOC. 1987, 109, 5355.
Lithium Dialkylamides and Related N-Lithiated Species Table I. Predicted NMR Resonance Multiplicities for 6Li-15NDoubly Labeled N-Lithiated Species structural form 6Li NMR spectrum 15NNMR spectrum R*N-//+LiS, singlet singlet R, R3~-~i
doublet
1:l:l triplet
Acc. Chem. Res., Vol. 26, NO.5, 1993 229
follow-up spectroscopic studies of [6Li,15N]LDA revealed only cyclic oligomer (presumed to be a dimer by analogy to LiCA) under all conditions and highlighted the limitation of structure determinations based solely on molality measurements that can be highly sensitive to impurities.
1:2:321 quintet R.. .R R&NOLi\Na>R
I
I
LI,
1:2:1 triplet
1:23:2:1 quintet
/LI N
47%.
R
R
Scheme I THF
THF
I
Phr,,
/Lk
Cy
\ Li 0
I
,,dy Ph
L \
I
I
THF
THF
Cy = 1cycbhexenyl
3
Ph>N-Li(THF)3
CY
4
within each class of aggregation state. Trisolvated monomers and tetrasolvated dimers having tetrahedral coordination about lithium show substantially smaller QSCs than do their trigonal planar disolvated monomer and disolvated dimer counterparts. As one might expect, higher solvation numbers correlate with strong donor solvents and sterically unhindered substrates. Although their method for solvation number determination has so far been limited to ArXLi (X= NR, 0) derivatives, it constitutes one of the most incisive solutions to the extremely difficult and persistent problem of lithium ion solvation state determination.'* Lithium Dialkylamides. Lithium isopropylcyclohexylamide (LiCA) is a commonly used base in organic synthesis. We turned to LiCA rather than more symmetrical lithium dialkylamides to capitalize upon the stereochemical consequences of aggregation.lg A single 13C NMR spectrum of LiCA evidenced aggregation by displaying twice the number of resonances anticipated for a simple monomer. PLi,I5N1LiCA in THF exhibited spectroscopicproperties fully consistent with a 1:l mixture of dimer stereoisomers at all LiCA and THF concentrations (eq 3). The depiction of LiCA dimers as disolvates is based upon indirect kinetic,20 theoretical,21 and s p e c t r o s c o p i ~evidence ~ ~ ~ ~ ~(vide ~~~ infra). Failure to observe any monomer was surprising in light of colligative measurements suggesting that a 0.05 M solution of lithium diisopropylamide (LDA) in neat THF contains primarily monomer.23 However, (18) For leading citations to lithium ion solvation state determination, see ref 21. (19) Galiano-Roth, A. S.; Michaelides, E. M.; Collum, D. B. J. Am. Chem. SOC.1988,110, 2658. (20) Galiano-Roth,A. S.;Collum, D. B. J. Am. Chem. SOC.1989,111, 6772. (21) Romesberg, F. E.; Collum, D. B. J. Am. Chem. SOC.1992, 114, n..
..
ZllZ.
(22) Romesberg, F. E.; Gilchrist, J. H.; Harrison, A. T.; Fuller, D. J.; Collum, D. B. J . Am. Chem. SOC.1991,113, 5751. (23) Seebach, D.; Bauer, W. Helu. Chim. Acta 1984,67, 1972.
The 6Li-15N double labeling method readily confirmed previous suggestions7that unsolvated lithium amides can exist in a variety of cyclic oligomers. Hexane solutions of solvent-free L6Li,l5N1LDA,for example, contain 3-5 discrete cyclic oligomers.24 In contrast, suggestionsthat lithium dialkylamidesin donor solvents exist only as cyclic dimers to the exclusion of higher cyclic oligomers were difficult to confirm for highly symmetric ca~es.~5 Moreover, the ambiguities stemming from the seemingly innocuous distinction of D2h dimers from Dnhhigher oligomers (see Table I) haunted rate and mechanism s t ~ d i e s . ~ ~ ~ ~ 6 ~ ~ Inverse-detected 15N homonuclear zero-quantum NMR spectroscopy provided a rigorous distinction of D2h cyclic dimers from Dnh higher oligomers. In the 15Nzero-quantum experiment, homonuclear 15Ntwospin coherence is prepared from the two 15N spins neighboring a 6Li atom in a 6Li-15N doubly labeled lithium dialkylamide cyclic oligomer.28 During the evolution period, the zero-quantum coherence will evolve under scalar coupling only to 6Lispins that couple to one, but not to both, 15Nspins. For a lithium amide dimer, all 6Lispins coupled to 15Nspins involved in the two-spin coherence couple to both 15N spins. Consequently, the coupling pattern will be a singlet along the fl dimension of the two-dimensional spectrum. In higher cyclic oligomers, there exist two 6Li spins that couple to one, but not both, 15N spins. The zeroquantum coherence will develop scalar coupling to the two nonshared 6Lispins, resulting in a 1:2:3:2:1 pattern along the fl dimension. As a result of a refocusing delay employed prior to detection, all oligomers will be a 1:2:l pattern along the f2 dimension. The 15Nzero-quantum NMR spectrum of PLi,l5N1LDA in THF (Figure 1A) shows no coupling in f1, consistent with a cyclic dimer.29 For comparison, the zero-quantum spectrum of lithium 2,2,6,6-tetramethylpiperidide (LiTMP) in benzene shows splitting in fl consistent with a cyclictrimer or tetramer (Figure 1B).B Other solvated lithium amide cyclic oligomers found to be cyclic dimers include LDA in TMEDA,27LiTMP in (24) Kim, Y.-J.; Bernstein, M. P.; Galiano-Roth,A. S.; Romesberg, F. E.; Williard, P. G.; Fuller, D. J.; Harrison, A. T.; Collum, D. B. J. Org. Chem. 1991,56,4435. (25) For a summary of the indirect evidence supporting the dimer assignments, see ref 22. 1988,110,5518. (26) Depue, J. S.; Collum, D. B. J. Am. Chem. SOC. Depue, J. S.; Collum, D. B. J. Am. Chem. SOC.1988, 110, 5524. (27) Bernstein, M. P.; Romesberg,F. E.; Fuller, D. J.; Harrison, A. T.; Collum, D. B.; Liu, Q.-Y.; Williard, P. G. J.Am. Chem. SOC. 1992,114, 5100. (28) Muller, L. J.Am. Chem. SOC.1979,101,4481. Bodenhausen, G.; Ruben, D. J. Chem. Phys. Lett. 1980,69,185. (29) Gilchrist, J. H.; Collum, D. B. J.Am. Chem. SOC.1992,114,794. ~~
Collum
230 Acc. Chem. Res., Vol. 26, No. 5, 1993
!E
5
E -5 8 W (12, Hz)
-10
10
5
0 -5 8 'U (12. Hz)
-18
Figure 1. 6Li-detected15Nzero-quantum NMR spectra of (A) 0.15 M [6Li,15N]LDA in THF at -90 O C and (B) 0.25 M [6Li,15N]LiTMPin 3:l benzene at 30 "C.
[GLi,l5N]LiHMDS confirmed the existence of a cyclic dimer in toluene, a mixture of oligomers in pentane (eq 2), and a monomer-dimer mixture in THF (eq 1).Thus, despite limited spectroscopictools available at the time, Kimura and Brown's assignmentsa are correct in every respect. Differentiation of the unsolvated cyclictrimers and tetramers must await further developments in multiple-quantum NMR spectroscopy. We began studying the solvation of lithium amides with N,N,iV',iV'-tetramethylethylenediamine (TMEDA) during the course of kinetic analyses of NJVdimethylhydrazone metalations by LDA. The spectroscopy clearly shows that LDA exists as a cyclic dimer in neat TMEDA. Moreover, a combination of spectroscopic, computational, and kinetic data support dimer 8 bearing labile (relative to THF) 1' linkages rather than the anticipated doubly chelated dimer 9.
THF,29LiHMDS in toluene,30and LiHMDS in THF.30 These results support the contention that solvated lithium amides in the presence of donor solvents are cyclic dimers rather than higher oligomers, whereas solvent-free lithium amides can exist as any of several D,,,cyclic oligomers. LiTMP is one of the most reactive and most hindered lithium amide bases commonly used to generate carbanions. In 1988, Renaud and Fox reported that the 'Li NMR spectra of LiTMP contain two resonances consistent with a monomer-dimer mixture.31 Followup studies of I6Li,l5N1LiTMP confirmed the existence of a monomer-dimer mixture in THF and revealed added subtleties due to conformational i s ~ m e r i s m . ~ ~ ~ ~ ~ 9 Although the single dimer resonance implicated the Me2N C2h conformer 6a rather than the CzVconformer 7a,we could not exclude the possibility that the facile (7.0 This result, taken in conjunction with an extensive kcal/mol) chair-chair ring flip was causing time aversurvey of the literature of TMEDA,33 led to the aging of the two signals. In contrast, conformationally suggestionthat TMEDA-mediated deaggregationsmay locked dimers 6b and 7b of lithium 2,2,4,6,6-pentabe more a consequence of the instability of the solvated methylpiperidide (LiPMP)32are forced to exchange by oligomers than the stability of the monomers. A a less facile ring scission process. Indeed, f6Li,l5N1preliminary test of this idea using [6Li,'5N]LiHMDS LiPMP displays a single 6Lidimer resonance in analogy proved provocative (Scheme 11). Treatment of lithium to [6Li,lSN]LiTMP. By slowing the site exchanges, hexamethyldisilazide (LiHMDS) with 5.0 equiv of THF differential N-Lidd and N-LiquaWrid coupling becomes affords exclusively dimer 10. LiHMDS in the presence readily discernible as a 6Li doublet of doublets. The of 5.0 equiv of TMEDA affords exclusively monomer observed coupling excludes any possibility of time 11. Contrary to a preponderance of conventional averaging of the two different resonances anticipated wisdom, a competition with equimolar quantities of for 7b. Thus, the high inherent symmetry typical of THF and TMEDA affords exclusively (THF-solvated) lithium amides is diminished in lithium piperidides by dimer rather than monomer. Thus, TMEDA competes conformational isomerism. Differential N-Li coupling poorly with THF for coordination sites on LiHMDS. and chemical shifts for axially and equatorially bound Moreover, the observable deaggregation in TMEDA/ lithium nuclei are readily observable in LiTMP mixed pentane does not necessarily result from a strongmetalaggregates as well (vide infra). ligand interaction.% We shall describe further evidence below that the highly revealing NMR spectroscopy of R = H; chairchair flip R = Me; N-Li bond scission lithium dialkylamides promises to shed light on issues I of organolithium aggregation and solvation in a general \ sense. The discussion of simple equilibria is concluded aptly with some exciting new results from the laboratory of Kenji Koga pertaining to the structures of chiral lithium amide~.3~ 6Liand 15NNMR spectroscopic analyses of PLi,15Nz]12 reveal solvent-dependent structure changes that appear to correlate with deprotonation enantioselectivities. Dimer 13 is the only observable form in toluene or etherholuene, while monomer 14 predomWith the key techniques in hand, we reinvestigated the structure of LiHMDSe30 Spectroscopic studies of (33) Collum, D.B. Ace. Chem. Res. 1992,25,448. li-N
-
L
(30) Romesberg, F. E.; Bernstein, M. P.; Fuller, D.J.; Harrison, A. T.; Collum, D.B. J. Am. Chem. SOC.,in press. (31) Renaud, P.; Fox, M. A. J . Am. Chem. SOC.1988,110, 5702. Harrison, A. T.;Fuller, D.J.; Collum, D. (32) Hall, P.; Gilchrist, J. H.; B. J . Am. Chem. SOC.1991,113,9575.
(34) For examples of poor correlationsof aggregationstate and solvent donicity, see: (a) Jackman, L. M.; Chen, X. J . Am. Chem. SOC.1992,114, 403. (b) Galiano-Roth, A. S.;Collum, D. B. J. Am. Chem. Soe. 1988,110, 3546. See also refs 22 and 27. (35) Sato,D.;Kawasaki,H.;Shimada,I.;Arata,Y.;Okamura,K.;Date, T.; Koga, K. J.Am. Chem. SOC.1992,114, 761.
Acc. Chem. Res., Vol. 26, No. 5, 1993 231
Lithium Dialkylamides and Related N-Lithiated Species Scheme I1
-niF
10
11
inates in THF, DME, or solvent mixtures containing HMPA. Of special note, the internal ligation by the pendant piperidine is retained in both 13 and 14 as evidenced by a small (2 Hz) coupling to the isotopically labeled piperidine moiety. The use of labeled amines to study internal ligation will certainly find applications in other areas of organolithium chemistry. The contrasting insights into chelation provided by lithium amide-TMEDA solvates and these chiral amides are also curious. The chelate effect continues to present interesting challenges to the mechanistically i n ~ l i n e d . ~
Scheme I11 THF
TAF
HMPA 0.5eqHMPA
I1
"
HMPA
HMPA 1 8
with other main group "ate" complexes38suggest that their relative obscurity is temporary. 12; X = N-piperiiino
13
%'
Ph
kPh 15
//
14
11. Complex Equilibria
HMPA-Solvated Lithium Amides. Despite its extreme carcinogenicity, hexamethylphosphoramide (HMPA; 15) remains the solvent of choice for enhancing organolithium reactivities; lithium amides are no ex~ e p t i o n .Recent ~~ investigations of the influence of HMPA on the structures of lithium amides have been especially illuminating. For example, Jackman and coworkers found that treatment of doublylabeled lithium anilides with HMPA affords mixtures of the anticipated monomers as well as triple ions of general structure 16.15 They made the astute observation that the factors often said to promote deaggregation also promote triple ion formation. Although triple ions are mentioned infrequently in the organolithium analogies (36) Fataftah, Z. A,; Kopka, I. E.; Rathke, M. W. J. Am. Chem. SOC. 1980,102,3959. Fraser, R. R.; Mansour, T. S. Tetrahedron Lett. 1986, 27, 331. Kodomari, M.; Sawa, S.; Morozumi, K.; Ohkita, T. Nippon Kagaku Kaishi 1976,301. Cuvigny, T.; Larcheveque, M.; Normant, H. Justw Liebigs Ann. Chem. 1975, 719. Reisdorf, D.; Normant, H. Organomet. Chem. Synth. 1972,1,375. Hoeomi, A,; Araki, Y.;Sakurai, H. J. Am. Chem. SOC.1982,104,2081. Cregge, R. J.;Hermann, J. L.; Lee, C. S.; Richman, J. E.; Schlessinger, R. H.Tetrahedron Lett. 1973,2425. Shirai, R.; Tanaka, M.; Koga, K. J . Am. Chem. Soc. 1986, 108, 543. Tsushima, K.; Araki, K.; Murai, A. Chem. Lett. 1989,1313. Tanaka, Y.; Tsujimoto, K.; Ohashi, M. Bull. Chem. SOC.Jpn. 1987,60, 788. (37) See ref 2 for an extensive bibliography of triple ions.
16
In 1989, the groups of Reich39 and Snaith40reported the first examples of two-bond Li-31P coupling by 6Li, 7Li, and 31P NMR spectroscopy, providing a view of lithium ion solvation of unparalleled clarity and importance. As applied to lithium amides, the 6Li-16N31Pcombination constitutes a highly effective triplelabeling method. For example, [6Li,16N]LDAin THF with added HMPA is shown to undergo a serial solvation of the dimer (Scheme 111)to give mixed solvate 17 and disolvate 18.22 Neither further solvation nor deaggregation occurs with excess HMPA. The asymmetry of 17 provided a means to exclude a cyclic trimer prior to the development of the 15Nzero-quantum NMR meth0d.29 The absence of a tetrasolvated dimer suggests that hindered amide dimers do not readily attain tetrahedral coordination at lithium. Surprised by the failure of HMPA to deaggregate LDA, we studied the influence of HMPA on the LiTMP (38) Richey, H. G., Jr.; Farkas, J., Jr. Organometallics 1990,9,1778. Fabicon, R. M.; Parvez, M.; Richey, H. G., Jr. J. Am. Chem. SOC.1991, 113,1412. Corey, E. J.; Hannon, F. J. Tetrahedron Lett. 1987,28,5233. Corriu. Pure Appl. Chem. 1988, 60, 99. Pelter, A.; Smith, K. In Comprehensive Organic Chemistry; Barton, D. H. R., Ollis, W.D., Us.; Pergamon Press: New York, 1979; Vol. 3, Chapter 14. Lipshutz, B. H. Synthesis 1987,325. Marko, I. E.; Rebiere, R. Tetrahedron Lett. 1992, 33, 1763. (39) Reich, H. J,; Green, D. P. J. Am. Chem. SOC.1989, 1 1 1 , 8729. Reich, H. J.; Green, D. P.; Phillips, N. H. J. Am. Chem. SOC.1989,111, 3444. Reich, H. J.; Borst, J. P. J. Am. Chem. SOC.1991, 113, 1835. (40) Barr, D.; Doyle, M. J.; Mulvey, R. E.; Raithby, P. R.; Reed, D.; Snaith, R.; Wright, D. S. J. Chem. SOC.,Chem. Commun. 1989, 318.
Collum
232 Ace. Chem. Res., Vol. 26, No. 5, 1993 Scheme IV
I
THF
h HM'PA
51
HM'PA
19
20
33a
,
,
, ,",
1.0
,
,
, ,
,
, , ,
0.0
.5
(, , .
,
-.5
-1.0
,
, ,
IP" 44
PPI(
Figure 2. +jLi-l5NH M Q C spectrum of 0.2 M [6Li,15N]LiHMDS w i t h 0.3 equiv of HMPA/Li in 3:l THF/pentane at -115 "C. 21
22
Scheme V
23
TH F
LfF
kN @Li-N
I 28
I THF
HMPA' 'HMPA
24
THF
[&-Li-Ny]-
25; *Li(HMPA)3(THF) 26; *Li(HMPA)4
monomer-dimer equilibrium.22 We surmised that addition of HMPA would shift the equilibrium completely to monomer. In fact, addition of HMPA to [6Li,16N]LiTMPin THF/pentane affords the bewildering distribution of species shown in Scheme IV. Tetrasolvated dimer 23, disolvated monomer 22, and tetrasolvated triple ion 26 are the observable forms in the limit of excess HMPA. Such oddities as open dimer 24, triple ions 25 and 26, and tetrasoluated dimer 23 are especially noteworthy. (Recall that the analogous tetrasolvate of dimeric LDA is not observable.) Most importantly, contrary to our original supposition, HMPA does not subs tantia 1ly deaggrega t e LiTMP. The failure of HMPA to deaggregate LDA or LiTMP prompted us to investigate its influence on the monomer-dimer mixture of LiHMDS in THF solution.30The overall structure assignments are illustrated in Scheme V. The parallels of LiHMDS and LiTMP in THF/ HMPA mixtures are striking. The primary difference is that triple ion 33a is observed in place of a LiHMDS open dimer analogous to 24. Once again, we find that HMPA does not appreciably deaggregate LiHMDS. Although the assignments in Scheme V were discerned from one-dimensional NMR spectroscopic methods with the aid to single frequency decouplings, the need for a more efficient two-dimensional analog was evident. 6Li-15N heteronuclear multiple quantum correlation (HMQC) spectroscopy4l affords the 6Li15N resonance correlations in a single experiment (Figure 2). The pulse sequence is the same as that (41)fiLi-16Nheteronuclear multiple quantum correlation spectroscopy was first evaluated using LiTMP-LiX mixed aggregates that had been characterized previously: Gilchrist, J. H.; Harrison, A. T.; Fuller, D. J.; Collum, D. B. Magn. Reson. Chem. 1992,30,855.
27
/THF
-Li/
Me3Si9N Me3Si
--Li
Me3Si
'HMPA
HMPA
'HMPA
29
30
THF
HMPA
Me3Siur, HLk ,stSiMe3 M e 3 S i d N\Li0NbSiMe3
Me3Siu,,, O L k .,&iMe3 Me3SivN\Li0N rSiMe3
I
I
I
I
1;;:
HMPA
HMPA
31
32
N-Li-NbSiMe3 ,..SiMe3
1-
33a;+Li(HMPA)2(THF)2, 33b;+Li(HMPA)3THF,and 33c;+Li(HMPA)d
used by Giinther and co-workersfor 6Li-13C correlation spectr~scopy.~~ Improved resolution and decreased overall spectrometer time provided substantial advantages relative to the one-dimensional methods. The analogous 6Li-31P HMQC spectroscopy was used to determine 6Li-31P resonance correlation^.^^**^ Lithium Amide-LiX Mixed Aggregates. It is now fully established that lithium salts generated during the course of an organolithium reaction can profoundly (42)Moskau, D.; Brauers, F.; Gijnther, H.; Maercker, A. J. Am. Chem. SOC.1987,109,5532.Gais,H.-J.;Vollhardt, J.;Ghther,H.;Moskau,D.;
1988,110,978.Giinther, H.; Lindner, H. J.; Braun, S. J.Am. Chem. SOC. Moskau, D.; Bast, P.; Schmalz, D. Angew. Chem.,Int. Ed. Engl. 1987,26, 1212. (43)For Li-P resonance correlations derived from single frequency decoupling, see: Denmark, S. E.; Miller, P. C.; Wilson,S. R.J.Am. Chem. SOC.1991,113, 1468.
Lithium Dialkylamides and Related N-Lithiated Species influence structure and reactivity by virtue of mixed aggregateformation.'* However, mixed aggregateshave been well characterized on only a few occasions.44 The 'jLi-15N double labeling methods offer opportunities to investigate mixed aggregation in detail. Kinetic studies of lithium diphenylamide (PhzNLi) N-alkylation revealed evidenceof autocatalysis by LiBr, prompting us to investigate the structure of f6Li,l5N1PhzNLi with added [6Li]LiBr.26At low THF concentration, ['jLi,l5N1PhzNLiexists as a disolvated dimer, 34. Addition of 1.0 equiv of i6Li1LiBr affords mixed dimer 35 quantitatively. The N-Li-N connectivity of 35 can be gleaned from the 6Li-15Ncoupling patterns. Rapid THF exchange precludes direct spectroscopic detection of two discrete lithium resonances for 35;the unsymmetrical solvation is based on colligative titration data. The placement of the bromide ion is based on analogy with the homonuclear amide dimers. At elevated THF concentrations, both dimer 34and mixed dimer 36 appear to be converted to THF-solvated monomer 36. Unfortunately, loss of 6Li-15N coupling renders the monomer assignment tentative.45 InterTHF
I
Phr~l,, RLi\ ,,,tPh PhrN\LiONbPh
I
THF
.
THF ;;2N-Li(THF),
3s
34
THF
Ace. Chem. Res., Vol. 26, No. 5, 1993 233
correlation. [6Li,15NlLDAfails to form mixed aggregates with unhindered lithium enolates such as PLi3 cyclohexenolate and forms mixed aggregates with hindered enolates reluctantly. However, [6Li,15N] LDA readily forms both 2:l and 1:l mixed aggregates with r6Li1LiC1and PLiILiBr (eq 4).47 The transannular Li-
Li-
;C#Li
X = CI, Br, but not OR
LI
X = CI, Br, and OR
- -X interaction in the 2:l mixed aggregate is depicted with a dotted line since ladders and cyclic trimers seem equally plausible. Ladders receive support from crystallographic a n a l ~ g ywhile , ~ trimers are supported by semiempirical (MNDO) computational studie~.~E Preliminary investigations of the factors influencing mixed aggregationwith LDA prove to be more interesting than enlightening.49 Within the isostmctural series 37a-d, the propensity to form LDA-LiXCMe3 mixed dimer follows the order X = NH > CHZ > 0 > S. Mixed aggregates bearing potentially chelating ligands such as 38a and 38b are not detectable, presumably because of disproportionate stabilization (by chelation) of the LiX homonuclear aggregate.3a
THF
35
estingly, all three structural forms were found to be kinetically reactive toward n-butyl bromide. In 1984, Corey and Gross reported that remarkable improvements in EIZ enolization selectivities are obtained when an extremely hindered lithium di-tertalkylamide base is used in the presence of chlorotrimethylsilane to trap the enolate as it forms.& Although explanations for EIZ enolization selectivities are typically couched in terms of competing thermodynamic and kinetic control, we suspected that mixed aggregation effects might be important. We addressed the influence of lithium enolates and lithium halides on the structure and reactivity of lithium amides using a combination of methods. Stereochemical studies of 3-pentanone enolization by LDA4' uncovered little evidence of intervening LDA-lithium enolate mixed aggregates. In contrast, a substantial influence of lithium halides on the EIZ enolization selectivity was detected. Parallel spectroscopic studies of mixed aggregation revealed an apparent structure-reactivity (44) Jackman, L. M.; Rakiewicz, E. F. J . Am. Chem. SOC.1991, 113, 1202 and references cited therein. (45) More recently it has been shown that treatment of PhnNLi in T H F with 4.0 equiv of HMPA affords exclusively [Ph,NLiNPh,l-//+Li(HMPA), (triple ion) rather than monomer. Romesberg, F. E.;Collum,D. B. Unpublished. Conductivity studies of PhaNLiin dilute THF solution are consistent with a contact ion pair. Krom, J. A.; Petty, J. T.; Streitwieser, A. Submitted for publication. (46) Corey, E. J.; Gross, A. W. Tetrahedron Lett. 1984,25, 491, 495. (47) Galiano-Roth, A. S.;Kim, Y.-J.;Gilchrist, J. H.; Harrison, A. T.; Fuller, D. J.; Collum, D. B. J. Am. Chem. SOC.1991, 113, 5053.
37a; X = NH 37b; X = CH2 37c; x = s 37d; x = 0
38a;S = NMe2 38b;S = OM8
Enolization stereochemical studies using LiTMP suggested that both LiTMP-lithium halide-and LiTMP-lithium enolate mixed aggregates may strongly influence LiTMP reactivityam Spectroscopic studies of [6Li,15N]LiTMPin the presence of lithium halides and lithium enolates revealed mixed aggregates of varying stoichiometry and conformational preference (Scheme VI).32 Although both lithium enolates and lithium halides (LiBr and LiC1) afford the 2:l and 1:l mixed aggregates (e.g., 39-41 and 42,respectively), only lithium cyclohexenolate forms a detectable 2:2 mixed aggregate (e.g., 43). The hindered enolate derived from diisopropyl ketone, on the other hand, affords a 1:l mixture aggregate to the exclusion of all other forms. Conclusions Organolithium reagents display a rich structural chemistry that is proving to be as complex and interesting as that observed for any of the transition elements. The lithium amides offer an excellent combination of synthetic utility, high kinetic reactivity, (48) Romesberg, F. E.; Collum, D. B. Unpublished. (49) Kim, Y.-J.; Collum, D. B. Unpublished. (50) Hall, P.; Gilchrist, J. H.; Collum, D. B. J. Am. Chem. SOC.1991, 113,9571. (51) Glassman, T. E.; Liu, A. H.; Schrock, R. R. I n o g . Chem. 1991,30, 4723.
234 Ace. Chem. Res., Vol. 26,No. 5, 1993
Collum
Scheme VI
40
41
and thermal stability. The 6Li-15N double labeling study offers a view of metal ion solvation and aggregation that is beginning to impact on areas beyond
traditional organolithium ~hemistry.~'As organolithium chemistry emerges as a legitimate subdiscipline of coordination chemistry, longstanding empirically determined relationships of aggregation state and solvent donicity are beginning to be replaced by working models founded upon firmly established organolithium solution structures. One of the goals of ongoing rate and mechanism studies is to more precisely define the role of solvation and aggregation in determining organolithium reactivity. I hope to present a detailed Account of lithium dialkylamide structure-reactivity relationships in the not-so-distant future. l a m deeply indebted to my talented co-workers for their skill, enthusiasm, and dedication. I a m also grateful to colleagues and to Aidan Harrison and David Fuller of our N M R facility for stimulating discussions and valuable contributions. T h e work was supported by the National Institutes of Health. I also acknowledge the National Science Foundation Instrumentation Program (CHE 7904,825 and PCM 8018643), the National Institutes of Health (RR02002), and IBM for support of the Cornell Nuclear Magnetic Resonance Facility.
